|Home | About | Journals | Submit | Contact Us | Français|
Lung cancer is the number one cause of cancer related deaths. The lack of specific and accurate tools for early diagnosis and minimal targeted therapeutics both contribute to poor outcomes. The recent discovery of microRNAs (miRNAs) revealed a novel mechanism for post-transcriptional regulation in cancer and has created new opportunities for the development of diagnostics, prognostics and targeted therapeutics. In lung cancer, miRNA expression profiles distinguish histological subtypes, predict chemotherapeutic response and are associated with prognosis, metastasis and survival. Furthermore, miRNAs circulate in body fluids and hence may serve as important biomarkers for early diagnosis or stratify patients for personalized therapeutic strategies. Here, we provide an overview of the miRNAs implicated in lung cancer, with an emphasis on their clinical utility.
Despite significant advances in detection and therapy, lung cancer remains the number one cause of cancer related deaths in both men and women. Unfortunately, most cases of lung cancer are detected at an advanced stage when the cancer has already metastasized thus limiting the chances for cure. Currently, low-dose computed tomography (LDCT) screening is employed in high risk patients to detect early stage disease. While this technique was shown to reduce mortality its usefulness is limited by high false-positive rates, exposure to potentially harmful radiation and no way to distinguish indolent nodules from tumors[1, 2]. These issues highlight the urgent need for accurate biomarkers that can detect early lung cancer with high sensitivity and specificity. In addition, investigators continue to recognize the molecular complexities of lung cancer and the value of high-throughput interrogation of the lung tumor genome in identifying novel actionable targets. For example, this approach has been successful in the development of therapeutics targeted towards tyrosine kinase receptor mutations and Anaplastic Lymphoid Kinase (ALK) rearrangements.
Since the discovery of microRNAs (miRNAs) in 1993 in C. elegans and in humans in 2000, nearly 2000 miRNAs have been reported accounting for 1–3% of human genes[4, 5]. MiRNAs are short (19–24 nucleotides) single-stranded RNAs predicted to regulate the expression of at least half of the human transcriptome controlling myriad biological processes including differentiation, proliferation, metabolism and apoptosis. A global reduction of mature miRNAs is observed in cancer, demonstrating the significance of maintaining tight regulatory control of miRNA biosynthesis to maintain cellular homeostasis. MiRNAs are initially transcribed by RNA polymerase II into a precursor primary miRNA (pri-miRNA), processed by the ribonuclease Drosha into a precursor hairpin (pre-miRNA) and exported out of the nucleus where they are cleaved by Dicer generating mature miRNA[7, 8]. They are then loaded into the RNA-induced silencing complex (RISC) together with Argonaute to silence target mRNAs. Decreased expression and/or mutations in these enzymes are associated with the development of lung tumors and are poor prognosis factor for lung cancer[10, 11].
In addition to functioning within cells, there is an abundance of miRNAs in body fluids exhibiting paracrine functions by targeting both the microenvironment and more distant sites in the body facilitating cell-cell communication. The identification of miRNAs in tissues and biological fluids including sputum, blood, plasma and urine has been used to classify several pathological conditions, including lung cancer[12–14]. MiRNA expression patterns permit an accurate discrimination between different histological subtypes and can identify the tissue of origin in cases of poorly differentiated tumors. Specific miRNA signatures from biological fluids thus have the potential to be useful non-invasive diagnostic and prognostic tools. Furthermore, because a single miRNA often targets multiple genes within a pathway, miRNAs are attractive targets for therapeutic intervention in cancer.
Here, we provide a review for the clinician focused on the clinical applicability of miRNAs in lung cancer, specifically as tools for diagnosis, prognosis and emerging targeted therapeutics.
Recently, several studies have demonstrated global dysregulation of miRNAs in lung cancer[15–17]. However, there remains a lack of consensus between these many studies, perhaps due to differences in sample type, preparation and the method of miRNA identification and analysis. Despite the issues of reproducibility, miRNA profiles from tissues may serve as important indicators and tools for classifying lung cancer subtypes and distinguishing primary from metastatic lung tumors (Table 1). For example, expression of miR-205 uniquely distinguishes squamous from non-squamous NSCLC even in poorly differentiated tumors[18–20]. This is of particular clinical relevance since screening for miR-205 provides a mechanism for distinguishing histological type in patients who have limited tissue available for diagnosis or in cases where the degree of differentiation and heterogeneity impedes diagnosis. Distinct miRNA signatures may also define tissue of origin. MiR-182 and miR-126 have characteristically opposing levels in primary versus lung metastasis from different organ sites. Additionally, miR-592 and miR-522 distinguished primary lung tumors from colon cancer metastasis in the lung.
The real potential for miRNAs, however, resides is in their presence and stability in biological fluids. Circulating miRNAs may reflect the tumor of origin and thus serve as novel noninvasive biomarkers for the early diagnosis of lung cancer and risk stratification of patients. A retrospective analysis of a miRNA signature in plasma from patients enrolled in the randomized (Multicenter Italian Lung Detection) MILD trial revealed a significant diagnostic performance for early detection of lung cancer[24, 25]. For their analysis, investigators developed a specific miRNA signature classifier (MSC) algorithm grouping patients into low, intermediate and high risk of cancer based on a predefined cutoff ratio value of 24 miRNAs identified from a previous training set. Each group was then examined for lung cancer occurrence, death and tumor stage. Diagnostic performance was compared between the MSC risk groups and LDCT. The MSC risk groups were associated with significantly different 3 year survival rates and had similar sensitivity and specificity as LDCT. However, combinations of the two techniques resulted in a five-fold reduction of LDCT false-positive rate. Thus, the authors concluded that MSC could complement LCDT screening. Additionally, a small set of plasma miRNAs differentiated between lung cancer and benign nodules offering clinicians a noninvasive approach to identify early lung cancers and avoid unnecessary surgery in patients with benign nodules. The clinical applicability of these miRNAs needs to be validated on a larger cohort but reinforces the use of circulating miRNAs for the diagnosis and risk stratification of patients with suspected lung cancer.
Lung cancers even when diagnosed at an early stage have a high incidence of recurrence compared to breast or prostate cancers . This startling fact speaks to the molecular heterogeneity that exists between early stage tumors. Preoperative and adjuvant chemotherapies provide only modest benefits for patients and come with many side effects, thus biomarkers to predict which patients may benefit the most from additional therapies are needed.
MiRNAs are associated with lung cancer driver mutations including EGFR, ALK and Ras as well as the tumor suppressors PTEN and p53, thus controlling numerous biological pathways driving tumor behavior (Table 2). Unfortunately, advances in targeted therapies such as EGFR tyrosine kinase inhibitor are only effective in patients harboring specific mutations and tumors often develop mutations resulting in a more aggressive, therapy resistant tumor [30, 31]. In the modern era of personalized therapy, identifying miRNAs that predict tumor aggressiveness and response to therapy could be used to better define prognosis and guide the clinician in determining the optimal personalized therapeutic approach.
The let-7 family of miRNAs was the first identified in humans and has since been implicated in Ras-driven lung cancer[5, 33]. Low let-7 levels are associated with a poor prognosis. Additionally, let-7 expression correlates with chemotherapeutic response and has been implicated in resistance[17, 36, 65–68]. Similarly, miR-34 and miR-449 are downregulated in lung cancer and under transcriptional control by p53 or E2F1, respectively[46, 48]. They are implicated in cell cycle control, apoptosis and invasion/migration via numerous targets namely CDK4, CDK6, MET, MYC and SIRT1. Other miRNA targets of E2F include the miR 15/16 family whose targets include cyclin D1, D2 and E1[47, 69].
MiR-128 expression is significantly downregulated in NSCLC correlating with pathological stage and metastasis. Furthermore, miR-128 regulates EGFR and miR-128 loss of heterozygosity correlated with clinical response to TKI. Other targets of miR-128 include VEGF and S6K1, thus the loss of miR-128 promotes proliferation and angiogenesis[40, 41].
Loss of miR-29 family members results in aberrant methylation patterns through induction of DNA methyltransferase (DNMT) 3A and 3B and re-expression restores normal patterns of DNA methylation inhibiting tumor formation in animal models. C-myc suppresses miR-29 expression contributing to FHIT promoter methylation and loss in lung cancer cells and patient tumors with low miR-29 expression had a poor prognosis. Additionally, expression of miR-29 inversely correlates with survival and through its interaction with ID1 might be predictive for the activity of Src inhibitors[43, 44].
Mir-34 is induced by p53 upon DNA damage or other oncogenic stresses[48, 70]. It is also regulated by methylation and thus is silenced in a number of cancer types, including lung cancer[71, 72]. Both the expression levels and methylation status of miR-34 have prognostic value in lung cancer[47, 70, 71]. Loss of miR-34 expression promotes proliferation and survival via target gene SNAIL, Met, MYC, HDAC1 and BCL-2[50, 51, 73]. Restoring miR-34 expression using mimics prevents cancer initiation and progression in mouse models of lung cancer and sensitizes lung cancer cells to radiation[74, 75].
EGFR is overexpressed in more than half of NSCLC and consequently is a target for directed therapeutic approaches such as antibodies and tyrosine kinase inhibitors (TKI). EGFR is a target of miR-7 and liposomal delivery of ectopic miR-7 sensitized EGFR+ lung cancer cells including EGFR-TKI-resistant lung cancer cells to TKI[42, 52].
The miR-200 family has been implicated in epithelial-mesenchymal transition (EMT) through its targets, ZEB1 and ZEB2 and E-cadherin. Because EMT is critical for metastasis and invasion and associated with chemoresistance it is not surprising that expression levels of miRNAs involved in this process, including mir-200 family members are prognostic markers in lung cancer[76, 77].
Clearly, tumor suppressive miRNAs control a wide variety of tumorigenic processes from proliferation and survival to metastasis providing strong rationale for developing therapeutic strategies that restore their expression.
Several miRNAs have been identified as functioning as oncogenes (oncomiRs).
High expression of miR-21 correlates with poor survival, recurrence and resistance in lung cancers [54, 78, 79]. Levels of miR-21 are higher in plasma from patients with an EGFR mutation than healthy patients or those without a mutation. Patients with higher expression of miR-21 had significant improvement in overall survival following adjuvant therapy with getfinib compared to those with low expression. Further, patient serum levels of miR-21 were significantly higher than baseline at the time of resistance to EGFR-TKIs suggesting that miR-21 may serve as a biomarker for acquired resistance. PTEN and PCD4 are two validated targets of miR-21 implicated in chemotherapeutic resistance.
MiR-141 is significantly upregulated in NSCLC and is a poor prognostic factor[76, 77]. Dysregulation suppresses the PI3K/Akt agonists PHLPP1 and PHLPP2 inducing proliferation and lung tumor growth. In ovarian cancer, miR-141 has been implicated in cisplatin resistance through suppression of KEAP1 and induction of EMT[80, 81].
MiR-221 and miR-222 suppress PTEN and TIMP3 promoting proliferation and survival and are implicated in TRAIL resistance in lung cancer. Additionally, EGFR and MET regulate the expression of both of these miRNAs and have important roles in EMT and getfitinib resistance through induction of BIM, and PKCε. These studies demonstrate the potential for miR-221 and miR-222 to serve as markers for therapeutic response and as potential therapeutic targets to sensitize tumors to TKI therapy.
The above studies demonstrate that alterations in miRNA expression may dictate cells response to therapy and thus could be early biomarkers of resistance. Numerous clinical trials are now ongoing to evaluate miRNA profiles that may stratify patients receiving a range of therapies, predict recurrence or metastasis or be used for surveillance following therapy. It is imperative that research continues to investigate the miRNAs that predict and overcome resistance in order to improve the options and outcomes for patients.
In addition to serving as diagnostic and prognostic tools to facilitate clinical decision making, miRNAs have the potential to serve as directed therapies themselves. The recent application of miRNAs as directed targets in humans is an encouraging sign of their potential as therapies for many human diseases. The initial results from these human based studies reveal the promise of miRNA targeted therapies. However, several obstacles must be overcome prior to miRNAs truly translating to the clinic. These obstacles include identifying the optimal methods for delivery, improving our understanding of pharmacokinetics of delivery, achieving cell specificity and minimizing off-target effects/toxicity. Currently, there are two main strategies for manipulating miRNAs as therapeutics in lung cancer, either restoring tumor-suppressor miRNA function or inhibiting the function of oncogenic miRNAs (Figure 1).
There are several methods employed to restore activity and function of tumor suppressive miRNAs including vector delivery of miRNA oligonucleotides or pharmaceuticals targeting global miRNA expression. MiRNA expression is regulated by transcription factors such as c-myc and p53, as well as through epigenetic modification via methylation. Thus, hypomethylating agents such as decitabine or 5-azacytidine may also be applied to reverse epigenetic silencing of miRNAs, however because these agents target all methylation it is unclear if effects are specific to miRNAs. These agents are approved for treating myelodysplastic syndromes, though their efficacy is limited by their nonspecific effects on global gene methylation.
A more specific approach to restoring miRNA function uses miRNA mimics reducing off-target effects and can be personalized based on the tumor’s miRNA signature. MiRNA mimics are synthetic double stranded RNAs with chemical modifications that improve stability and cellular uptake and are processed into a single-strand form to regulate gene expression similar to miRNAs. The guide strand is identical to the miRNA of interest and the passenger strand often contains chemical modifications such as cholesterol which enhance uptake or chemical modifications to prevent RISC loading.
Vector based delivery systems were first employed for gene therapy but have varied efficiency, stability, permeability, duration of genomic changes and off-target toxicities[86, 87]. MiRNA expression vectors are engineered with promoters that drive tissue or tumor-specific expression of the miRNA of interest. For example, adenoviral and lentiviral delivery of let-7 reduced lung tumor burden in a KRAS-driven mouse model of lung cancer[88–90]. Lipid emulsions, liposomes and nanoparticles have also been employed to improve the stability and uptake of miRNA mimics. Delivery of let-7 and miR-34 via the blood stream using a neutral lipid emulsion inhibited tumors in a KRAS-driven mouse model of lung cancer, thus demonstrating the potential clinical utility for this approach. Additionally, cationic liposomes have been used to effectively deliver miR-7 to EGFR-resistant xenograft tumors reducing tumor volume dramatically. Alternatively, nanoparticles can be coated with tumor specific antibodies thus enhancing the tumor specific effects of the mimics and reducing off targeting effects. Using a model of metastatic melanoma, investigators delivered miR-34 via nanoparticles coated with a tumor-targeting single-chain variable fragment (scFv) reducing lung metastasis.
These studies provide preclinical evidence for the potential translation of miRNA restoration techniques to treat lung cancer. MiRNA-34 loss is implicated in a number of cancers including hepatocellular carcinoma and lung cancer. Preclinical studies have demonstrated that restoration of miR-34 reduced tumor growth in animal models[93-95].The first miRNA replacement therapy, MRX34 is a liposome formulate miR-34 mimic that is intravenously injected. In 2013, MRX34 recently entered human clinical trials for patients with advanced or metastatic liver cancer (clinicaltrials.gov).
Since the discovery of oncogenic miRNAs, strategies for inhibiting them have been developed primarily based on anti-sense oligonucleotides. These approaches include antagomirs, locked nucleic acid (LNA) miRNAs and miRNA sponges (Figure 1). Antagomirs are synthetic RNAs complementary to the targeted miRNA blocking its function allowing mRNA to be translated. Antagomirs are limited for clinical application because large doses are required for effective miRNA blocking, though chemical modifications that enhance uptake, promote binding and prevent degradation improve this limitation[98, 99]. LNA anti-miRNAs substitute several nucleic acids with bicyclic RNA analogues holding the anti-miRNA in a locked confirmation. This modification permits shorter oligonucleotides with high affinity to their target miRNAs to be used with increasing potency and reducing off target effects. Similar to antagomirs, miRNA sponges bind mature miRNAs disrupting their function; however, they are encoded in a viral vector driven by a strong promoter and contain multiple tandem binding sites complementary to a heptamer in the miRNA of interest seed sequence permitting the sponge to block an entire miRNA seed family. The clinical applicability of miRNA sponges is similarly limited by the dose required to inhibit miRNA activity and off target effects from the viral vector insertion.
A variety of in vitro and in vivo studies have demonstrated the potential of targeting miRNAs using these approaches; however, only one has reached clinical trials. The first of its kind to enter human clinical trials, Miravirsen is a LNA-based therapeutic for the treatment of hepatitis C viral (HCV) infection[102, 103]. MiRNA-122 is a liver specific miRNA essential for stability and viral replication of hepatitis C virus. The LNA-anti-miR-122 binds and sequesters mir-122 preventing viral replication reducing viral titers in HCV-infected patients. The success of this therapeutic strategy in treating hepatitis C infection demonstrates its potential for applicability in treating other diseases including cancer.
Lung cancer is a complex disease complicated by a lack of tools for early detection and clinical decision making. The emerging role of miRNA control of gene networks and thus biological pathways associated with lung cancer is providing new hope for the identification of non-invasive and accurate miRNA signatures to diagnosis and guide therapeutic decisions for lung cancer patients which can lead to novel therapeutic targets.
While it is clear that miRNAs are differentially expressed in normal lung tissues versus lung disease (such as COPD) and lung cancer, the lack of reproducibility identifying specific miRNAs signatures has impeded the development of a reproducible miRNA signature for early detection. This lack of consensus supports the urgent need for the standardization of protocols for sample processing, miRNA detection and analysis to improve the identification of miRNA biomarkers. MiRNAs differ significantly between tissue and plasma and are dependent on sample processing, specifically contamination with platelets [105, 106]. Further complicating the identification of signatures are the multitude of platforms for identifying differentially expressed miRNAs including custom and commercial arrays based, solexa sequencing and RT-qPCR. In addition, there is a general lack of a consistency in data normalization and analysis. These will all need to be addressed in order to use miRNAs as an accurate noninvasive early detection tool for lung cancer.
Despite these obstacles, circulating miRNAs have the potential to change the landscape of patient care, providing a noninvasive and accurate method to identify miRNA signatures that in turn guide the design of a personalized therapeutic strategy. The stability of miRNAs in body fluids is attributable to an association to lipoproteins, RNA binding proteins or packaging within extracellular vesicles (EVs) like exosomes[107–109]. Recent evidence demonstrates that exosomes isolated from biological fluids may provide a more reliable and consistent miRNA profile than unseparated miRNA. Exosomes play a central role in cell communication through transfer of their contents.[108, 111]. Tumor derived exosomes are abundant in bodily fluid, contain increased concentrations of miRNAs that reflect the tumor and influence neighboring and distant cells thus making them attractive targets as biomarkers and novel therapeutics[108, 110–116].
Promising preclinical studies have shown that the use of miRNA mimics and anti-miRNAs can be successfully used to restore normal gene networks in animal models of lung cancer. However, implementing miRNA and anti-miRNA therapeutic strategies remains a significant hurdle for clinicians. Some concerns remain as to the best delivery method and stability of the miRNA therapies in the body as well as the safety profile of potential miRNA-targeted therapies. Further interrogation into the effects of chemical modifications, off-target gene regulation and duration of response is needed to move miRNA-targeted therapeutics into the clinic for treatment of lung cancer.
Additionally, because miRNAs target multiple genes simultaneously it is important to understand the implications of miRNA-targeted therapy especially when used in cooperation with other targeted therapies, chemotherapy or radiation. Some may have the potential to overcome resistance or to sensitize patients to therapies, but the possibility of activating an alternative mechanism for resistance also exists. It is therefore critical to evaluate the potential off-target effects of the therapeutic miRNAs as a single agent and in combination with current therapeutic regimens.
Despite these challenges, two miRNA based therapies have moved into the clinic with encouraging results. Though much work still remains to elucidate the specific miRNA(s) to target for lung cancer it is clear that the identification and validation of miRNA signatures will guide therapeutic decisions improving outcomes in the near future.
This work was supported by National Institutes of Health (NIH) grants 5R21CA179403-02 (SPN) and T32HL007946 (JFB).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.